three-address code


Also found in: Wikipedia.

three-address code

[′thrē ′ad‚res ‚kōd]
(computer science)
In computers, a multiple-address code which includes three addresses, usually two addresses from which data are taken and one address where the result is entered; location of the next instruction is not specified, and instructions are taken from storage in preassigned order.
Mentioned in ?
References in periodicals archive ?
We assume that the general form of statements in a program to be transformed is three-address code extended to include loop headers and array references.
The three-address code retains the loop headers and array references from the source program, which enables the user to specify loop-level transformations and array transformations.
The data-flow analysis phase performs global optimization on the three-address code. The rule base is consulted to determine if there are features on the target processor that allow optimization.